MOLECULAR AND CELLULAR BIOLOGY, Oct. 1990, p. 5333-5339 Vol. 10, No. 10 0270-7306/90/105333-07$02.00/0 Copyright C) 1990, American Society for Microbiology The Leucine Zipper of c-Myc Is Required for Full Inhibition of Erythroleukemia Differentiation MICHAEL J. SMITH,1 DENISE C. CHARRON-PROCHOWNIK,2 AND EDWARD V. PROCHOWNIKl 3* Division of HematologylOncology, Department ofPediatrics,' The Committee on Cellular and Molecular Biology,3 and The School ofPublic Health,2 The University of Michigan School of Medicine, Ann Arbor, Michigan 48109 Received 23 October 1989/Accepted 30 June 1990 The leucine zipper motif has been observed in a number of proteins thought to function as eucaryotic transcription factors. Mutation of the leucine zipper interferes with protein dimerization and DNA binding. We examined the effect of point mutations in the leucine zipper of c-Myc on its ability to dimerize in vitro and to inhibit Friend murine erythroleukemia (F-MEL) differentiation. Glutaraldehyde cross-linking studies failed to provide evidence for homodimerization of in vitro-synthesized c-Myc protein, although it was readily demonstrated for c-Jun. Nevertheless, whereas transfected wild-type c-myc sequences strongly inhibited F-MEL differentiation, those with single or multiple mutations in the leucine zipper were only partially effective in this regard. Since the leucine zipper domain of c-Myc is essential for its cooperative effect in ras oncogene-mediated transformation, this study emphasizes the close relationship that exists between transfor- mation and hematopoietic commitment and differentiation. c-Myc may produce its effects on F-MEL differentiation through leucine zipper-mediated heterodimeric associations rather than homodimeric ones. Several functions have been attributed to the c-myc proto- (20). Point mutagenesis of the leucine zipper of Fos and Jun oncogene. For example, its enforced expression from an family proteins has borne out this prediction and established inducible promoter in mitotically quiescent cells partially the importance of the dimer structure in DNA binding (12, mimics some of the features of growth factor stimulation (3). 13, 16, 24, 31, 32). High levels of exogenous c-myc potentiate the effects of A functional role for the c-Myc leucine zipper has been several types of growth factors (33). c-myc can cooperate suggested by Dang et al., who demonstrated that C-terminal with c-ras to transform some types of primary cells in vitro deletions of c-Myc, which abolished the leucine zipper (19) and in certain instances can act alone to produce such domain, also eliminated cooperativity with the EJ ras onco- changes (15). Finally, alterations in the levels of c-myc in the gene in the transformation of primary rat embryo fibroblasts Friend erythroleukemia and HL-60 promyelocytic leukemia (9). They also showed that bacterially expressed c-Myc cell lines exert profound effects on differentiation (8, 10, 14, protein could form homodimers as well as higher-order 18, 29, 40). oligomers. Whereas the biological effects imparted by c-myc are In this report, we have addressed the role of the c-Myc manifold, the mechanisms through which these effects occur leucine zipper in the inhibition of Friend murine erythroleu- are less well understood. Although it is widely believed that kemia (F-MEL) differentiation. In our initial studies, we c-myc is a transcriptional activator, the evidence to support attempted to duplicate the findings of Dang et al. by showing this notion, while compelling, is largely indirect and, in some that in vitro-translated c-Myc protein could form ho- cases, incompatible with this notion (27). Although v-Myc modimers. Despite the ease with which this was accom- and c-Myc proteins are capable of binding both single- plished for c-Jun protein, we were unsuccessful in demon- stranded and double-stranded DNA (1, 2, 6), no specific strating it in the case of c-Myc. Nevertheless, when target sequence has been identified as in the case of other expression plasmids containing point mutations in the c-Myc nuclear proto-oncogene products such as c-Fos, c-Jun, and leucine zipper were introduced into F-MEL cells, all muta- c-Myb (2, 4, 5). When fused at its N terminus to the tions showed a severe impairment in their ability to inhibit DNA-binding region of the LexA repressor in yeast cells, dimethyl sulfoxide (DMSO)-induced differentiation. These c-Myc can activate transcription of genes spaced down- reports suggest that any leucine zipper-mediated interactions stream of the LexA operator (21). However, this effect is of c-Myc which do occur are probably hetrodimeric in much less pronounced than with c-Fos, and its relevance to nature. transcriptional activation in higher cells is uncertain. c-Myc is also one of a number of proteins that contain a MATERIALS AND METHODS so-called leucine This zipper (20). structural motif consists of Cell culture and transfections. A subclone of the original a C-terminal region of high alpha-helical content with four to F-MEL 745 line was used in all experiments. In the absence five leucine residues evenly spaced at seven-amino-acid of chemical agents that induce erythroid differentiation, intervals. When displayed on an idealized helix wheel, the hydrophobic leucine side chains project from the alpha helix <1% of these cells routinely stained positive with the benzidine reagent (28). After a 5-day exposure to 1.5% in a common plane. It was originally predicted that this would allow for the interdigitation of the zipper regions of DMSO, however, >70% of the cells became benzidine positive. The cells were grown as previously described in opposing molecules and provide a means of dimerization Dulbecco modified Eagle minimal essential medium contain- ing 10% supplemented calf serum (Hyclone, Logan, Utah), 2 mM glutamine, 100 U of penicillin G per ml, and 100 ,ug of * Corresponding author. streptomycin per ml in a humidified 5% CO2 atmosphere 5333 5334 SMITH ET AL. MOL. CELL. BIOL. (28). Care was taken to maintain cells at a sufficiently low umn. TrpE-c-Myc fusion protein competed completely for density (generally <5 x 105/ml) to ensure logarithmic growth authentic c-Myc protein in Western immunoblot assays. at all times. Transfected cells were maintained under iden- Detection of c-Myc protein in F-MEL cells. F-MEL cells tical conditions except that the culture medium contained were either untreated or treated with 1.8% DMSO for 5 h. 10% dialyzed supplemented calf serum and 0.25 ,uM meth- Cells were pelleted by low-speed centrifugation, washed otrexate (MTX) (28, 29). Benzidine staining was performed three times in phosphate-buffered saline, and lysed in stan- as previously described (28, 29). Transfections were per- dard RIPA buffer containing 100 ,ug of phenylmethysulfonyl formed by using a Gene Pulser (Bio-Rad Laboratories, fluoride per ml and 1 ,ug each of aprotinin, leupeptin, and Richmond, Calif.) with settings of 1.0 kV and 25 ,uF and a pepstatin per ml. Approximately 100 ,ug of total protein per resultant time constant of 0.5 to 0.6 ms. A total of 2 x 107 ml was electrophoresed through a 7.5% polyacrylamide-SDS cells in 0.5 ml of phosphate-buffered saline were subjected to gel and then transferred to nitrocellulose by electroblotting. transfection, using 20 ,ug of linearized c-myc expression The filter was then blocked for 4 to 6 h in a 5% suspension of vector plasmid DNA and 2 p.g of linearized pSV2neo DNA nonfat dried milk prepared in TBS buffer (0.15 M; 20 mM (34). Successfully transfected clones were selected in me- Tris hydrochloride [pH 7.6]). Undiluted antibody was then dium containing G418 (0.48 mg/ml, active concentration; added to a final dilution of 1:1,000, and incubation was GIBCO Laboratories, Grand Island, N.Y.). Pooled G418- continued for 2 h at 25°C. After extensive washing ofthe blot resistant clones were then subsequently selected in medium in TBS plus 0.5% Tween 20, bound anti-c-myc antibody was containing 0.25 ,uM MTX and maintained under these con- detected with biotinylated goat anti-rabbit antibody, fol- ditions. Individual clones were obtained by limited dilution lowed by immunodetections with an avidin-alkaline phos- of the MTX-resistant cell population in 96-well microtiter phatase-based detection kit (Bethesda Research Laborato- plates. ries, Inc., Gaithersburg, Md.). Mutagenesis of murine c-myc cDNA. The starting template Glutaraldehyde cross-linking studies. Full-length murine in all c-jun cDNA was a kind gift from R. Bravo (EMBL Labora- cases was single-stranded M13mpl8 bacteriophage tories, Heidelberg, Federal Republic of Germany). The DNA containing a 2.0-kilobase HindlIl fragment or murine cDNA and full-length c-myc cDNA were cloned into SP6 c-myc cDNA (35). The single-stranded DNA was prepared expression vectors, transcribed in vitro, and capped, using a from phage grown in Escherichia coli CJ236 (dut ung thi Promega mRNA 5' capping kit according to the directions of relA) as described by Kunkel (17), using the Bio-Rad Muta- the supplier. After DNase digestion and phenol-chloroform Gene in vitro mutagenesis kit. Specific 20-nucleotide (nt)- extraction, 1 ,ug of each RNA was translated in vitro in 50 ,ul long oligomers containing single nucleotide mismatches to of rabbit reticulocyte lysate containing 50 ,uCi of [355] each of the leucine zipper codons were used to generate methionine (specific activity, approximately 1,200 ,uCi/ point mutations. The identity of each mutant was confirmed mmol; Amersham Corp., Arlington Heights, Ill.). Cross- by direct DNA sequencing, using a primer complementary to linking was carried out essentially as described by Nak- a region of cDNA just upstream from that encoding amino abeppu et al.